Browsing by Author "Pournaki, Armin"

Now showing 1 - 4 of 4

Grounding force-directed network layouts with latent space models
(2023) Gaisbauer, Felix; Pournaki, Armin; Banisch, Sven; Olbrich, Eckehard
Force-directed layout algorithms are ubiquitously used tools for network visualization. However, existing algorithms either lack clear interpretation, or they are based on techniques of dimensionality reduction which simply seek to preserve network-immanent topological features, such as geodesic distance. We propose an alternative layout algorithm. The forces of the algorithm are derived from latent space models, which assume that the probability of nodes forming a tie depends on their distance in an unobserved latent space. As opposed to previous approaches, this grounds the algorithm in a plausible interaction mechanism. The forces infer positions which maximise the likelihood of the given network under the latent space model. We implement these forces for unweighted, multi-tie, and weighted networks. We then showcase the algorithm by applying it to Facebook friendship, and Twitter follower and retweet networks; we also explore the possibility of visualizing data traditionally not seen as network data, such as survey data. Comparison to existing layout algorithms reveals that node groups are placed in similar configurations, while said algorithms show a stronger intra-cluster separation of nodes, as well as a tendency to separate clusters more strongly in multi-tie networks, such as Twitter retweet networks.
How Influencers and Multipliers Drive Polarization and Issue Alignment on Twitter/X - Data (Version v1) [Data set]
(2025) Pournaki, Armin; Gaisbauer, Felix; Olbrich, Eckehard
We provide anonymized retweet networks extracted from trending topics in Germany collected between 2021 to 2023. More specifically, we collected tweets from 2021-03-29 to 2023-07-12 according to the following scheme: at the beginning of each day, we launched a script that collects the current "trending topics" (from now on referred to as "trends") in Germany using the Twitter Trend API (v1). By default, trends are personalized based on the account's Twitter/X usage. One can, however, disable the personalization by setting a specific location from which to draw the trending topics, which then yields "popular topics among people in a specific geographic location" (X/Twitter2025). We re-ran the script every 15 minutes. At the end of each day, we counted the number of times each trending topic appeared during the day and kept the top 5 most frequent ones. This gave us a proxy of the five most important trending topics for that day. We then used the Twitter Search API (v1) to collect German-speaking tweets using the exact trend keyword as a query on the day it trended and the day after (48hrs). All the tweets were collected using a single Twitter API key, collecting tweets for maximally 24 hours every day. For each trend, we extract a retweet network, in which nodes are Twitter users and a directed link is drawn from user to if retweets . We provide one retweet network for each trend as a csv after anonymizing the user_ids. There is one csv for each trend containing the columns source,target,weight. The filename contains the date and the keyword that was searched: T__.csv All the individual files are contained in rtn.zip. Additionally, we computed a topic model on the full text of tweets which allowed us to classify each trend into one larger metatopic (such as Covid, Climate Change, Sports, ...). This topic assignment is contained in trend2topic.csv. For more information on the topic model, please refer to the paper https://doi.org/10.1609/icwsm.v19i1.35890.
Ideological differences in engagement in public debate on Twitter
(2021) Gaisbauer, Felix; Pournaki, Armin; Banisch, Sven; Olbrich, Eckehard; Guidi, Barbara
This article analyses public debate on Twitter via network representations of retweets and replies. We argue that tweets observable on Twitter have both a direct and mediated effect on the perception of public opinion. Through the interplay of the two networks, it is possible to identify potentially misleading representations of public opinion on the platform. The method is employed to observe public debate about two events: The Saxon state elections and violent riots in the city of Leipzig in 2019. We show that in both cases, (i) different opinion groups exhibit different propensities to get involved in debate, and therefore have unequal impact on public opinion. Users retweeting far-right parties and politicians are significantly more active, hence their positions are disproportionately visible. (ii) Said users act significantly more confrontational in the sense that they reply mostly to users from different groups, while the contrary is not the case.
Twitter Explorer: A Framework for Observing Twitter through Interactive Networks
(2021) Pournaki, Armin; Gaisbauer, Felix; Banisch, Sven; Olbrich, Eckehard
We present an open-source interface for scientists to explore Twitter data through interactive network visualizations. Combining data collection, transformation and visualization in one easily accessible framework, the twitter explorer connects distant and close reading of Twitter data through the interactive exploration of interaction networks and semantic networks. By lowering the technological barriers of data-driven research, it aims to attract researchers from various disciplinary backgrounds and facilitates new perspectives in the thriving field of computational social science.